课程地址:https://class.coursera.org/nlp/lecture/31
1. What is Sentiment Analysis?
情感分析(Sentiment analysis)又可以叫做
意见抽取(Opinion extraction)
意见挖掘(Opinion mining)
情感挖掘(Sentiment mining)
主观分析(Subjectivity analysis)等等。
比如对电影好坏的评价:
比如Google Product Search 上面对产品属性的评价,并展示褒贬程度:
还比如Bing Shopping对上面同一台打印机的评价结果:
Twitter sentiment versus Gallup Poll of Consumer Confidence
通过Tewitter上和民意调查得来的数据的进行对比,发现两者对于消费者信心的统计结果有很大的相关性(correlation),相关度达到80%。
同样,利用Twitter上的公众情绪预测股票,发现:CLAM(平静)情绪的指数可以预测三天后道琼斯的指数。
Target Sentiment on Twitter(Twitter Sentiment App):用户可以输入产品关键字来查询其他网友对此产品的评价
Why sentiment analysis?
- Movie: is this review positive or negative?
- Products: what do people think about the new iPhone?
- Public sentiment: how is consumer confidence? Is despair increasing?
- Politics: what do people think about this candidate or issue?
- Prediction: predict election outcomes or market trends from sentiment
电影产品预测,民意政治倾向等等…
Sentiment analysis is the detection of attitudes
分析用户对人或物品的态度(attitudes:enduring, affectively colored beliefs, dispositions towards objects or persons)
分析主体包括:
- Holder (source) of attitude :谁有这个属性
- Target (aspect) of attitude :评价的对象
- Type of attitude :有哪些方面
- From a set of types
Like, love, hate, value, desire, etc. - Or (more commonly) simple weighted polarity:
positive, negative, neutral, together with strength
- From a set of types
- Text containing the attitude :文本是单句还是整句
Sentence or entire document
由简单到复杂的情感分析:
- Simplest task:
- Is the attitude of this text positive or negative? 简单的极性
- More complex:
- Rank the attitude of this text from 1 to 5 评分等级
- Advanced:
- Detect the target, source, or complex attitude types 对象,来源及复杂的态度等。